Will an AI take your job? Diving into ChatGPT’s learning ecosystem

The topic that’s sparked countless conversations coming into 2023 is ChatGPT. These seven letters now appear on every social media platform, people are obsessing over it across blog posts, and the L&D industry in particular is dying to know more about it.

ChatGPT is a learning model produced by OpenAI, a leading laboratory for Artificial Intelligence research and production, founded in late 2015. Their mission is ‘…to ensure that artificial general intelligence benefits all of humanity.’ They focus on autonomous systems that can outperform humans in economically-valuable work, ensuring the AI is effective, but also safe. There are people that doubt the rise in AI technology as a safe route that can always be in human hands, and there are those that doubt the actual success of such an endeavour.

Let’s take a look and dive into the world of Artificial Intelligence, to see what makes ChatGPT tick.

What does ChatGPT do?

ChatGPT is a machine learning model that understands ‘natural’ language in text – something which hasn’t been seen on this scale before. It’s designed to work like a human brain, identifying data patterns using interconnected neurons, which helps it to predict what comes next in text. OpenAI trained it on data from the internet as well as human trainers, all focusing on human conversations to achieve its natural-sounding tone of voice. Mike Sharples, a professor of Educational Technology at the Open University UK, alludes to the fact ChatGPT is ‘…creating an internal representation, not just at the surface text, but of the ideas and concepts behind it.’

The ultimate goal of this AI model is to communicate and create content, all while appearing as human as possible. So far, we’ve seen a variety of well-written articles and natural-sounding conversations as a result of its interactive help chat. People across the globe have reported the model reaching maximum capacity on numerous occasions, with a ‘get notified when we’re back page’, proving that its popularity and usefulness are not going away any time soon:

There’s no doubt about what ChatGPT aims to do and how well it does it. What we want to know is how this Artificial Intelligence model is somehow replicating what millions of human beings do every day.

How does it do this?

ChatGPT is based on OpenAI’s original GPT-3 model, a language prediction model. Its neural network ultimately absorbs text from a human and uses it to transform it into an answer that the AI predicts will be the most useful, which is based on more input information. The GPT-3 is fed vast amounts of online information and communication to identify and recognise patterns, which is how the AI builds its answers. ChatGPT is an upgraded version of this original learning model, bringing new levels of accuracy and effectiveness in a broader range of applications. Even though the genius behind all of this is all human, the implications AI might have on humans is not necessarily in our hands.

What happens to people’s jobs?

For many, the first thing that comes to mind is, if ChatGPT can write such great articles, what will happen to journalists, content writers, and every profession that falls under this umbrella? The main thing that sets us apart from AI, which could ultimately save our jobs, is the difference in emotional integrity. For example, if you were to ask ChatGPT for advice on how to commit suicide, the AI would give you accurate advice based solely on research and logic. Human beings are the only ones to respond emotionally and urge you not to do something so extreme. Emotional integrity is important in any kind of writing, whether it’s a romantic novel or social media post because it will be emotional humans reading and reacting it, not AI. 

There are positives coming from ChatGPT, and AI in general, when it comes to training and teaching. From the classroom to the office, Artificial Intelligence can aid both the educator and the learner with efficiency and punctuality, as it doesn’t require physical breaks and rest like human beings. However, a core element of L&D departments in particular is the role of the coach. Can ChatGPT replace a human coach? The same question as before arises with the need for emotional depth, particularly when well-being at work is concerned. However, for on-the-job training, ChatGPT can certainly help to an extent. Its only limitations would be the specific interactions required for people with learning difficulties and other disabilities not so easily defined by humans, let alone a logic-based AI. ChatGPT and other AIs replacing human coaches in the workplace would be an interesting process, and Saffron Interactive explored this in one of our LSG webinars, presenting an actual AI coach called ‘AIDA’ (Artificially Intelligent Development Assistance). 

People in IT and other technological professions can rejoice! As soon as you have skills in IT, you’re all set for a future possibly dominated by AI. The biggest surge in upskilling will arguably be the technological and digital skills needed for repairing AI, managing AI, and basically all things AI. The issue arises when we think of the pressure around upskilling and reskilling. The sheer amount of people that will enter the industry, along with the people that want to stay, will possibly be through the roof. In theory, this could pose an increase in the need for more coaches, which could both be human and AI, so theoretically, it could benefit on all fronts. However, the pressure placed on organisations to make this happen could be catastrophic. The growth for the IT and tech industry could be huge, but could also be equally stressful for every L&D professional involved. 

Introducing ChatGPT to L&D, or the other way around?

How does this all relate to the future of learning and development? Workplace learning has largely been supported by technology over the last couple of years, especially with the pandemic imposing remote working, and ChatGPT could just be another development in this field. There have been L&D technological advances that have tried to benefit from the pandemic and haven’t quite hit the mark they hoped for, such as gamification. However, lots of people seem convinced that OpenAI will be a name spoken about for a long time.

ChatGPT can certainly help people learn, this has been proven with its instantaneous help chat and writing talents integrated with Grammarly. But the question is how beneficial is it in helping people learn using methods we don’t already have? The immediate contender for this is interactive learning, especially for answering employees’ questions quickly and efficiently. ChatGPT responds instantly with only factual answers, so this is a huge help to organisations with a large workforce, and many new employees.

Donald H Taylor, Chair of the Learning Technologies Conference, expressed his insights into ChatGPT as a learning model. He pointed out why this logic-based model still makes mistakes, sharing a conversation he had with ChatGPT. Donald said that ‘It has no intent. It is a machine. It is a machine doing what it is programmed to do. Moral: do not confuse fluency with accuracy…’

Written for OEB Global 2023 by Chloë Sibley.